Zhang Rou, Liu Zhijuan, Li Ran, Ai Li, Li Yongxia
Department of Respiratory Medicine and Critical Care Medicine, The Second Affiliated Hospital of Kunming Medical University, Kunming, People's Republic of China.
Nat Sci Sleep. 2025 May 24;17:1049-1066. doi: 10.2147/NSS.S520758. eCollection 2025.
Pulmonary hypertension (PH) is a common cardiovascular complication of obstructive sleep apnea (OSA), posing a significant threat to the health and life of patients with OSA. However, no clinical prediction model is currently available to evaluate the risk of PH in OSA patients. This study aimed to develop and validate a nomogram for predicting PH risk in OSA patients.
We collected medical records of OSA patients diagnosed by polysomnography (PSG) from January 2016 to June 2024. Transthoracic echocardiography (TTE) was performed to evaluate PH. A total of 511 OSA patients were randomly divided into training and validation sets for model development and validation. Potential predictive factors were initially screened using univariate logistic regression and Lasso regression. Independent predictive factors for PH risk were identified via multivariate logistic regression, and a nomogram model was constructed. Model performance was assessed in terms of discrimination, calibration, and clinical applicability.
Eight independent predictive factors were identified: age, recent pulmonary infection, coronary atherosclerotic heart disease (CHD), apnea-hypopnea index (AHI), mean arterial oxygen saturation (MSaO), lowest arterial oxygen saturation (LSaO), alpha-hydroxybutyrate dehydrogenase (α-HBDH), and fibrinogen (FIB). The nomogram model demonstrated good discriminative ability (AUC = 0.867 in the training set, AUC = 0.849 in the validation set). Calibration curves and decision curve analysis (DCA) also indicated good performance. Based on this model, a web-based nomogram tool was developed.
We developed and validated a stable and practical web-based nomogram for predicting the probability of PH in OSA patients, aiding clinicians in identifying high-risk patients for early diagnosis and treatment.
肺动脉高压（PH）是阻塞性睡眠呼吸暂停（OSA）常见的心血管并发症，对OSA患者的健康和生命构成重大威胁。然而，目前尚无临床预测模型可用于评估OSA患者发生PH的风险。本研究旨在建立并验证一种用于预测OSA患者PH风险的列线图。
我们收集了2016年1月至2024年6月通过多导睡眠图（PSG）诊断的OSA患者的病历。进行经胸超声心动图（TTE）以评估PH。共511例OSA患者被随机分为训练集和验证集，用于模型的开发和验证。首先使用单因素逻辑回归和Lasso回归初步筛选潜在预测因素。通过多因素逻辑回归确定PH风险的独立预测因素，并构建列线图模型。从区分度、校准度和临床适用性方面评估模型性能。
确定了8个独立预测因素：年龄、近期肺部感染、冠状动脉粥样硬化性心脏病（CHD）、呼吸暂停低通气指数（AHI）、平均动脉血氧饱和度（MSaO）、最低动脉血氧饱和度（LSaO）、α-羟丁酸脱氢酶（α-HBDH）和纤维蛋白原（FIB）。列线图模型显示出良好的区分能力（训练集AUC = 0.867，验证集AUC = 0.849）。校准曲线和决策曲线分析（DCA）也表明模型性能良好。基于该模型，开发了一个基于网络的列线图工具。
我们建立并验证了一种稳定且实用的基于网络的列线图，用于预测OSA患者发生PH的概率，有助于临床医生识别高危患者以便早期诊断和治疗。
